The Identification And Verification Of Species-Specific MicroRNAs And Their Pre-micrornas In Lampetra Japonica

MicroRNAs(miRNAs)has been shown to regulate the expression of many genes involved in various bio-logical processes of metabolism after transcription.In recent years,the discovery and function analysis of miRNA have been extensively studied in the animal world with the application of deep sequencing and microarray technology.In the current study,the small RNAs of Lamprey(Lampetra japonica)leucocytes were conducted high-throughput sequencing by using the second generation sequencing technology.A total of 5207787 small RNA sequences were identified,and 4739346 were identified as miRNA and assembled to 10989 variants.The 10989 miRNA variants belong to 306 highly conserved miRNA family according to sequence similarity based on the known miRNAs.In addition,the other 70 se-quences were predicted to be new miRNAs and their precursor were also predicted.Thirty-four of the newly predicted miRNAs were verified to be expressed in leucocytes of L.japonica treated with antigen by miRNA microarray assay.Among them,the minimal folding free energy indexs of 16 pre-miRNAs were exceed 0.85.These results indicate that the specific miRNA exist in L.japonica,and these regulatory miRNAs might play an important role in lamprey leukocyte growth,development and in the response to disease.Further validation of high-throughput sequencing technologies of the proceeds of the new the authenticity of the miRNA sequences,we also introduce support vector machine(SVM)method.Using lamprey miRNA as the input vectors of support vector machine tool Lib-SVM,we develop a program Lj-miRSVM for the predication of lamprey precursor miRNA sequence after optimizing parameters with the Grid program.Through the 10-fold cross validation method,we obtain that the average sensitivity,specificity,accuracy are 95%,95.5%,95.85,respectively,and the area ratio under the ROC is 98.9%.We developed the forecast the application feasibility of lampreys is very good.At the same time,we predict the pre-miRNA of other 20 species using the Lj-miRSVM classifier and obtain 78.5% average accuracy.Illustrate the classification efficiency of the process is good.